Upload folder using huggingface_hub
Browse files- .gitattributes +5 -0
- cyberagent/open-calm-3b/Constant_251_attr__value +3 -0
- cyberagent/open-calm-3b/Constant_265_attr__value +3 -0
- cyberagent/open-calm-3b/_gpt_neox_layers.0_attention_rotary_emb_Constant_5_attr__value +0 -0
- cyberagent/open-calm-3b/_gpt_neox_layers.0_attention_rotary_emb_Constant_attr__value +0 -0
- cyberagent/open-calm-3b/config.json +28 -0
- cyberagent/open-calm-3b/generation_config.json +6 -0
- cyberagent/open-calm-3b/onnx/decoder_model.onnx +3 -0
- cyberagent/open-calm-3b/onnx/decoder_model.onnx_data +3 -0
- cyberagent/open-calm-3b/onnx/decoder_model_merged.onnx +3 -0
- cyberagent/open-calm-3b/onnx/decoder_model_merged.onnx_data +3 -0
- cyberagent/open-calm-3b/onnx/decoder_model_merged_quantized.onnx +3 -0
- cyberagent/open-calm-3b/onnx/decoder_model_quantized.onnx +3 -0
- cyberagent/open-calm-3b/onnx/decoder_with_past_model.onnx +3 -0
- cyberagent/open-calm-3b/onnx/decoder_with_past_model.onnx_data +3 -0
- cyberagent/open-calm-3b/onnx/decoder_with_past_model_quantized.onnx +3 -0
- cyberagent/open-calm-3b/quantize_config.json +103 -0
- cyberagent/open-calm-3b/special_tokens_map.json +6 -0
- cyberagent/open-calm-3b/tokenizer.json +0 -0
- cyberagent/open-calm-3b/tokenizer_config.json +10 -0
.gitattributes
CHANGED
@@ -45,3 +45,8 @@ cyberagent/open-calm-1b/_gpt_neox_layers.0_attention_rotary_emb_Constant_attr__v
|
|
45 |
cyberagent/open-calm-1b/onnx/decoder_model.onnx_data filter=lfs diff=lfs merge=lfs -text
|
46 |
cyberagent/open-calm-1b/onnx/decoder_model_merged.onnx_data filter=lfs diff=lfs merge=lfs -text
|
47 |
cyberagent/open-calm-1b/onnx/decoder_with_past_model.onnx_data filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
45 |
cyberagent/open-calm-1b/onnx/decoder_model.onnx_data filter=lfs diff=lfs merge=lfs -text
|
46 |
cyberagent/open-calm-1b/onnx/decoder_model_merged.onnx_data filter=lfs diff=lfs merge=lfs -text
|
47 |
cyberagent/open-calm-1b/onnx/decoder_with_past_model.onnx_data filter=lfs diff=lfs merge=lfs -text
|
48 |
+
cyberagent/open-calm-3b/Constant_251_attr__value filter=lfs diff=lfs merge=lfs -text
|
49 |
+
cyberagent/open-calm-3b/Constant_265_attr__value filter=lfs diff=lfs merge=lfs -text
|
50 |
+
cyberagent/open-calm-3b/onnx/decoder_model.onnx_data filter=lfs diff=lfs merge=lfs -text
|
51 |
+
cyberagent/open-calm-3b/onnx/decoder_model_merged.onnx_data filter=lfs diff=lfs merge=lfs -text
|
52 |
+
cyberagent/open-calm-3b/onnx/decoder_with_past_model.onnx_data filter=lfs diff=lfs merge=lfs -text
|
cyberagent/open-calm-3b/Constant_251_attr__value
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:83dd5f1a5a07912d2e7dbb73b75edeb2ce4c1f1d6ec48144a887bd4a4f72325b
|
3 |
+
size 4194304
|
cyberagent/open-calm-3b/Constant_265_attr__value
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:83dd5f1a5a07912d2e7dbb73b75edeb2ce4c1f1d6ec48144a887bd4a4f72325b
|
3 |
+
size 4194304
|
cyberagent/open-calm-3b/_gpt_neox_layers.0_attention_rotary_emb_Constant_5_attr__value
ADDED
Binary file (655 kB). View file
|
|
cyberagent/open-calm-3b/_gpt_neox_layers.0_attention_rotary_emb_Constant_attr__value
ADDED
Binary file (655 kB). View file
|
|
cyberagent/open-calm-3b/config.json
ADDED
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "cyberagent/open-calm-3b",
|
3 |
+
"architectures": [
|
4 |
+
"GPTNeoXForCausalLM"
|
5 |
+
],
|
6 |
+
"attention_dropout": 0.0,
|
7 |
+
"bos_token_id": 0,
|
8 |
+
"classifier_dropout": 0.1,
|
9 |
+
"eos_token_id": 0,
|
10 |
+
"hidden_act": "gelu",
|
11 |
+
"hidden_dropout": 0.0,
|
12 |
+
"hidden_size": 2560,
|
13 |
+
"initializer_range": 0.02,
|
14 |
+
"intermediate_size": 10240,
|
15 |
+
"layer_norm_eps": 1e-05,
|
16 |
+
"max_position_embeddings": 2048,
|
17 |
+
"model_type": "gpt_neox",
|
18 |
+
"num_attention_heads": 32,
|
19 |
+
"num_hidden_layers": 32,
|
20 |
+
"rope_scaling": null,
|
21 |
+
"rotary_emb_base": 10000,
|
22 |
+
"rotary_pct": 1.0,
|
23 |
+
"tie_word_embeddings": false,
|
24 |
+
"transformers_version": "4.33.2",
|
25 |
+
"use_cache": true,
|
26 |
+
"use_parallel_residual": false,
|
27 |
+
"vocab_size": 52224
|
28 |
+
}
|
cyberagent/open-calm-3b/generation_config.json
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 0,
|
4 |
+
"eos_token_id": 0,
|
5 |
+
"transformers_version": "4.33.2"
|
6 |
+
}
|
cyberagent/open-calm-3b/onnx/decoder_model.onnx
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:00dc656852abedbb1dbe4c053f84375297fac7e91704ec2a3da9f6745fa0847b
|
3 |
+
size 7527647
|
cyberagent/open-calm-3b/onnx/decoder_model.onnx_data
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c78496ffcc5c963d521413df42036e6255f0216608b9f7db7d26da3933a1a564
|
3 |
+
size 11140157440
|
cyberagent/open-calm-3b/onnx/decoder_model_merged.onnx
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:330fd712ef92498111380973c9ac5384b511e2bc67b1c6161522c04b8629f8f6
|
3 |
+
size 15068313
|
cyberagent/open-calm-3b/onnx/decoder_model_merged.onnx_data
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c78496ffcc5c963d521413df42036e6255f0216608b9f7db7d26da3933a1a564
|
3 |
+
size 11140157440
|
cyberagent/open-calm-3b/onnx/decoder_model_merged_quantized.onnx
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8045695874443b09704db8eb10ca452488c17a66cab320e752eaa0fece193f50
|
3 |
+
size 2808808466
|
cyberagent/open-calm-3b/onnx/decoder_model_quantized.onnx
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ce5f90f7287f10d5c84069a6bee66a3877410a86326fa457ca6103da105eae61
|
3 |
+
size 2800480061
|
cyberagent/open-calm-3b/onnx/decoder_with_past_model.onnx
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:154d146e69339f57cf58e9ff4f4b4886c175601d2b50f295f6cd7ca3533777d9
|
3 |
+
size 7573186
|
cyberagent/open-calm-3b/onnx/decoder_with_past_model.onnx_data
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c78496ffcc5c963d521413df42036e6255f0216608b9f7db7d26da3933a1a564
|
3 |
+
size 11140157440
|
cyberagent/open-calm-3b/onnx/decoder_with_past_model_quantized.onnx
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:23cf9d6be9621ea4cb4179e39e42d194f3b028e9e87d8c2b2701836025262f94
|
3 |
+
size 2800533153
|
cyberagent/open-calm-3b/quantize_config.json
ADDED
@@ -0,0 +1,103 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"per_channel": true,
|
3 |
+
"reduce_range": true,
|
4 |
+
"per_model_config": {
|
5 |
+
"decoder_model_merged": {
|
6 |
+
"op_types": [
|
7 |
+
"Mul",
|
8 |
+
"Softmax",
|
9 |
+
"Erf",
|
10 |
+
"Where",
|
11 |
+
"Cast",
|
12 |
+
"Shape",
|
13 |
+
"Tile",
|
14 |
+
"Range",
|
15 |
+
"Unsqueeze",
|
16 |
+
"Constant",
|
17 |
+
"GatherElements",
|
18 |
+
"If",
|
19 |
+
"Concat",
|
20 |
+
"Neg",
|
21 |
+
"Add",
|
22 |
+
"Transpose",
|
23 |
+
"Pow",
|
24 |
+
"Expand",
|
25 |
+
"Reshape",
|
26 |
+
"Sub",
|
27 |
+
"Gather",
|
28 |
+
"Div",
|
29 |
+
"ConstantOfShape",
|
30 |
+
"MatMul",
|
31 |
+
"ReduceMean",
|
32 |
+
"Sqrt",
|
33 |
+
"Slice",
|
34 |
+
"Squeeze"
|
35 |
+
],
|
36 |
+
"weight_type": "QInt8"
|
37 |
+
},
|
38 |
+
"decoder_model": {
|
39 |
+
"op_types": [
|
40 |
+
"Mul",
|
41 |
+
"Softmax",
|
42 |
+
"Erf",
|
43 |
+
"Where",
|
44 |
+
"Cast",
|
45 |
+
"Shape",
|
46 |
+
"Tile",
|
47 |
+
"Range",
|
48 |
+
"Unsqueeze",
|
49 |
+
"Constant",
|
50 |
+
"GatherElements",
|
51 |
+
"Concat",
|
52 |
+
"Neg",
|
53 |
+
"Add",
|
54 |
+
"Transpose",
|
55 |
+
"Pow",
|
56 |
+
"Expand",
|
57 |
+
"Reshape",
|
58 |
+
"Sub",
|
59 |
+
"Gather",
|
60 |
+
"Div",
|
61 |
+
"ConstantOfShape",
|
62 |
+
"MatMul",
|
63 |
+
"ReduceMean",
|
64 |
+
"Sqrt",
|
65 |
+
"Slice",
|
66 |
+
"Squeeze"
|
67 |
+
],
|
68 |
+
"weight_type": "QInt8"
|
69 |
+
},
|
70 |
+
"decoder_with_past_model": {
|
71 |
+
"op_types": [
|
72 |
+
"Mul",
|
73 |
+
"Softmax",
|
74 |
+
"Erf",
|
75 |
+
"Where",
|
76 |
+
"Cast",
|
77 |
+
"Shape",
|
78 |
+
"Tile",
|
79 |
+
"Range",
|
80 |
+
"Unsqueeze",
|
81 |
+
"Constant",
|
82 |
+
"GatherElements",
|
83 |
+
"Concat",
|
84 |
+
"Neg",
|
85 |
+
"Add",
|
86 |
+
"Transpose",
|
87 |
+
"Pow",
|
88 |
+
"Expand",
|
89 |
+
"Reshape",
|
90 |
+
"Sub",
|
91 |
+
"Gather",
|
92 |
+
"Div",
|
93 |
+
"ConstantOfShape",
|
94 |
+
"MatMul",
|
95 |
+
"ReduceMean",
|
96 |
+
"Sqrt",
|
97 |
+
"Slice",
|
98 |
+
"Squeeze"
|
99 |
+
],
|
100 |
+
"weight_type": "QInt8"
|
101 |
+
}
|
102 |
+
}
|
103 |
+
}
|
cyberagent/open-calm-3b/special_tokens_map.json
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": "<|endoftext|>",
|
3 |
+
"eos_token": "<|endoftext|>",
|
4 |
+
"pad_token": "<|padding|>",
|
5 |
+
"unk_token": "<|endoftext|>"
|
6 |
+
}
|
cyberagent/open-calm-3b/tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
cyberagent/open-calm-3b/tokenizer_config.json
ADDED
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_prefix_space": false,
|
3 |
+
"bos_token": "<|endoftext|>",
|
4 |
+
"clean_up_tokenization_spaces": true,
|
5 |
+
"eos_token": "<|endoftext|>",
|
6 |
+
"model_max_length": 1000000000000000019884624838656,
|
7 |
+
"pad_token": "<|padding|>",
|
8 |
+
"tokenizer_class": "GPTNeoXTokenizer",
|
9 |
+
"unk_token": "<|endoftext|>"
|
10 |
+
}
|